Will Card Sorting determine top level categories

Process & Techniques
#1

Hi,

I’m hoping for some advice/feedback. I recently had a client pushback on a sitemap, saying they felt some secondary pages deserved top-level positioning as opposed to being grouped into a main category. I chose that particular main category based on a strategy I formed during our discovery process, but I can see the client’s side as well.

Is there a process that will help me gain insight into top-level categories? I’m assuming that card sorting already has these in place, but I could be wrong. Please give me your thoughts… I’d appreciate your expertise.

-J

1 Like
#2

hi @jeff_doubek

IMO you are fully right.
I believe that the card-sorting is the best exercise/tool to find a compromise with your client.
I often run workshops with them using such tool to make them aware of the complexity level and to highlight, business-wise, what are the most important task to be accomplished.
This makes easier for them to join the other steps of the design process, they will have more background and will be more in sync with your deliverables

3 Likes
#4

Looks interesting. Are you affiliated with the website?

#6

A “little” late but to clarify for others looking at card sorting in research. Card sorting will help with defining categories, your information architecture, and uncovering audience mental models. Specifically, there are a number of different card sort techniques that will help with different problems:

  • Open Card Sort: Participants define the groups/categories which helps you discover their mental models and map organization knowledge to user expectations.
  • Reverse Card Sort (also called a tree sort or inverse sort): Users try to find an item in an existing hierarchy. Good for testing an existing set of categories/info architecture. I think this might of helped the most in your situation, it’s relatively quick to test multiple proposed info architectures with this method.
  • Modified Delphi Sort: A little advanced, but participants build on each others’ groupings. I’ve had a good experience with these, but wouldn’t recommend unless you have some experience. (White paper)
  • Closed Card Sort: You predefine the group labels and ask participants to put cards in them. Usually not recommended because you are only lightly testing your own labels and not uncovering user insights. Most of the time you are better off using an open sort or a reverse sort. There are specific use cases a closed sort might be the right choice, but always reconsider one of the other methods first. I’ve used closed sorts for existing taxonomy research but won’t for information architecture.

Cheers!